Until OptimalTest combined adaptive testing techniques with …
- Modern database technology
- Real-time software tools
- Seamless networking
|
|
… many viewed adaptive testing only as a way to reduce test time. But with OptimalTest’s
comprehensive, innovative test management solutions, adaptive testing also:
- Significantly improves yield learning, yield reclamation, reliability
and quality Achieves these improvements in real time so that data becomes immediately
actionable so products and operations can evolve.
- Processes more data with more automation and higher data integrity
while enhancing its value as useful, accessible information.
Modern testing has progressed beyond separating good chips from bad. The multiple
disciplines of design, fab, test, quality and reliability all put unique demands
on the effectiveness of the test program. In the past, these demands were prohibited
by cost and the lack of real-time data to differentiate one chip from another. Therefore,
traditional testing applied the same, compromised test program to all chips. However
OptimalTest’s application of adaptive testing encompasses enterprise-wide networks,
modern databases and the advantages of real-time results -- as well as two innovative
methodologies for establishing reference dies and flexible rule sets – to unleash
the power of leveraging data across the enterprise.
Current Test Methods
OptimalTest Methods
OptimalTest’s innovations in adaptive testing techniques give you the latest generation
of test management tools for significant, continual improvements in
- Yield learning for better design and manufacturing processes across
the entire IC lifecycle
- Yield reclamation, test time reduction, reliability, quality
- Lowering test costs and increasing test capacity
OptimalTest’s implementation of “intelligent” adaptive testing goes beyond the common
view of adaptive test as simply skipping tests to reduce test time. Instead, OptimalTest
- Adds tests to targeted devices to give you greater
- yield learning and yield, up to 5% at wafer sort and final test
- reliability and quality on ALL devices while reducing test times on reference
dies by up to 30%
- Focuses on data integrity and data accessibility
- Leverages recent advances in data management and analysis tools
- Provides a new standard of process control
The innovative technology that enables OptimalTest’s advanced adaptive testing includes
modern database technology combined with adaptive testing. Among its benefits:
- Providing data you can act on
- The ability to establish reference dies/units
- A consolidated, open algorithm (rule) engine
OptimalTest’s unique, updated approach for intelligent adaptive testing requires
that control and discipline be maintained via two critical means -- reference dies/units
and test rules -- with patented technology for establishing reference dies/units
for all wafers, lots and products. This capability allows you to immediately create
yield and reliability baselines, giving you essential information for making further
decisions regarding yield, test time reduction, reliability and quality while simultaneously
realizing reduced costs and other business benefits. Decisions about test time reduction
and reliability can be made in real time and decisions from yield learning, offline.
Test Program Standard vs. Optimal
For example, on the wafer, OptimalTest software selects locations of reference dies
according to various algorithms that consider their locations on the die and other
important information. They are tested first with the full test program to assure
quality and monitor the health of the tester/lot/wafer. Then, based on the results,
test augmentation or test reduction will be executed. Moreover, tests for yield
learning and reliability will be always executed for enhanced baseline on the selected
reference die.
We’ve also developed a unique capability to create, simulate and apply test rules.
Our customizable, open algorithm engine enables you to control the test process
and establish the environment for state-of-the-art adaptive testing. It enables
rapid, simple generation of real-time and off-line rules for simulation (using historical
test data) or execution. This sophisticated approach, incorporating auto-learning
and auto execution, greatly contributes to increases in yield, reliability and quality.
In addition, test time can be reduced without changing the test program, with adaptive
TTR controlled by a rule set external to the program and completely transparent
to operations. (
Also see Consolidated Open
Algorithms Engine.)